INDEX

Explanations

prefixes or words with the letters "im" followed by a single consonant

oai_token-act-pair · gpt-3.5-turbo

expressions of personal feelings and states of being

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ridge

-0.73

rower

-0.71

 Christy

-0.70

yards

-0.67

Coy

-0.66

 Slate

-0.66

 Chao

-0.66

anche

-0.64

 Chavez

-0.64

YC

-0.63

POSITIVE LOGITS

im

3.73

Im

1.81

IM

1.45

Im

1.38

hes

1.35

imm

1.15

 imbalance

1.14

 imitation

1.10

 Imam

1.08

1.05

Activations Density 0.013%

prefixes or words with the letters "im" followed by a single consonant

expressions of personal feelings and states of being

No Comments

No Known Activations